After much debate about slowing AI development, Anthropic and OpenAI are releasing new models. Both iterate on Fable 5.1 and GPT-6 Astra, offering similar performance at lower costs.

Anthropic's Opus 5.5 is better at complex work, software inefficiencies, financial analysis, and business tasks—targeting enterprise customers. Benchmarks claim it beats GPT-6 Astra on Terminal-Bench 4.0 and Frontier Code v1.1 (Main). It produces clearer writing and "attempted to circumvent boundaries around 85 percent less often." Pricing is $4 input/$20 output per token, down from $5/$25.

OpenAI's GPT-6 Sol and GPT-6 Luna focus on efficiency. They improve professional work, coding, and computer use, up to 50% cheaper than GPT-5.6 promotional pricing. Sol costs $2 input/$10 output; Luna $0.10 input/$0.50 output. Both are better at facts—Sol makes half as many mistakes—and match Fable 5.1 in coding. They communicate clearer and respond faster via prompt caching. OpenAI says they lie less about coding results and deny more bypass attempts.

Opus 5.5 is available via Claude, AWS, Google Cloud, and Microsoft Azure. GPT-6 Sol and Luna are in ChatGPT Work and Codex for Plus, Pro, Business, and Enterprise; free and Go users get only Luna on desktop.